Water distribution networks (WDNs) are essential infrastructure that should be planned to take into account the uncertain future in which they will operate. According to Creaco et al. (2021), nodal demands are critical inputs for modeling WDNs and their variability is one of the main sources of uncertainty that affects network sizing (Magini et al., 2019). For this reason, several researchers propose that WDN design should be developed taking several possible demand conditions into consideration. But when many demand scenarios are used, design definition becomes more complex. Powerful tools are needed to find optimal designs, particularly when more than one objective is pursued. Taking multiple demand scenarios into account can avoid the outcome being an under-or over-sized network, which can happen when only a single under-or over-estimate of demand is made.The expansion of smart water meter technology at service connections in WDNs in recent years has made many high-frequency measurements of water consumption available. The literature is rich in heterogeneous consumption data sets at various spatial scales, ranging from city district to a single household or individual water fixture, and several temporal sampling frequencies, from monthly up to sub-daily: hour, minute, or second (Di Mauro et al., 2021). The availability of so much data supports planners and water utility managers in choosing the best design solution and operational strategy but, also shows up the high degree of variability that marks water demand, mainly due to the unpredictable behavior of human beings. Statistical and data mining tools support descriptive and predictive analytics, which can capture demand variability at different scales in space and time, derive statistical moments, define appropriate probability density functions, or model suitable stochastic processes.There is a large body of literature dealing with the "statistical uncertainty" (Walker et al., 2003) of water demand. In some cases, the main purpose of the studies is to analyse demand, for example, for time pattern recognition, end-use disaggregation, simulation, and forecasting. In other cases, they consider water demand uncertainty when addressing WDN design and management issues, such as for optimal network sizing (Salcedo-Díaz et al., 2020),